fps-automated-test

Calculates in an automated way, with selenium, the frames per second for an HTML page. It uses the Javascript requestAnimFrame, in order to calculate the framerate variation. Please refer to the file: src/javascript/README.md for more insight on the Javascript part

Disclaimer

The extracted fps values are not the same as the ones you may experience during your normal user experience.(in the same environment)

Never the less the extracted values are a good measure as long they satisfy the following principles:

• the extracted values are consistent - we have the same mean and the same std deviation between different runs with the same input
• it shows a degrading in performance under the same input as when a normal user experiences the same

The performance is consistent and could vary more or less from what a normal user is experiencing but still it will be a variation in the same direction

In order to better match your user experience it is required to modify the fps.py file The code does not give any guaranty of your actual performance, it just gives a relative measure of it.

Requirements:

pip install -r requirements.txt

• python
• selenium
• lettuce
• numpy

Before run:

• Please make sure you have all the required dependencies installed:
• Also please insert all the values inside: perf_util.py

How to run it:

After you have made the changes required to match your page, you can run it with the lettuce command in the src folder.

Case study:

Twitter -> look into the example folder to see how the script is modified to check the Twitter performance.

Running the script, you will not see any degradation in performance between 100 and 500 scrolls, because there is none.

Thanks:

The javascript is inspired from the original work of mrdoob: https://github.com/mrdoob/stats.js

Many thanks